Handbook of Statistical Analysis and Data Mining Applications

Chapter 20

Top 10 Data Mining Mistakes

OUTLINE

Preamble

Introduction

0 Lack Data

1 Focus on Training

2 Rely on One Technique

3 Ask the Wrong Question

4 Listen (Only) to the Data

5 Accept Leaks from the Future

6 Discount Pesky Cases

7 Extrapolate

8 Answer Every Inquiry

9 Sample Casually

10 Believe the Best Model

How Shall We Then Succeed?

Postscript

Preamble

Mining data to extract useful and enduring patterns remains a skill arguably more art than science. Pressure enhances the appeal of early apparent results, but it is all too easy to fool yourself. How can you resist the siren songs of the data and maintain an analysis discipline that will lead to robust results? It is essential to not lack (proper) data, focus on training, rely on one ...